Downscaling the variable sfcWind

Results

cesm2.ssp245

<<<<<<< HEAD

<<<<<<< HEAD ======= >>>>>>> scaler <<<<<<< HEAD ======= >>>>>>> scaler

Metric	lstm	xgboost	naive	cnn	cnn	lstm	xgboost	naive
diff_of_means	1.026	1.534	1.575	-1.621
ratio_of_sd	0.946	0.909	0.914	0.949
amplitude_ratio_of_means	0.768	0.609	0.641	0.740
maximum_error	0.190	0.154	0.173	0.138
ks_mean_on_coarse_res_with_extremes	0.186	0.347	0.323	0.259
qqplot_mae	0.056	0.093	0.103	0.087
acf_mae	0.087	0.140	0.126	0.100
extremogram_mae	0.033	0.074	0.065	0.054
-0.838	1.370	1.534	1.575
ratio_of_sd	0.980	0.969	0.909	0.914
amplitude_ratio_of_means	0.816	0.779	0.609	0.641
maximum_error	0.208	0.148	0.154	0.173
ks_mean_on_coarse_res_with_extremes	0.196	0.179	0.347	0.323
qqplot_mae	0.048	0.058	0.093	0.103
acf_mae	0.087	0.094	0.140	0.126
extremogram_mae	0.043	0.034	0.074	0.065

Time series of the first days

Plots

Important: Right now we are only estimating the upper tail extremogram. Currently we didn’t find a way to estimate the two tales at the same time. We are using quant = .97

<<<<<<< HEAD

=======

>>>>>>> scaler

cesm2.ssp370

<<<<<<< HEAD

<<<<<<< HEAD ======= >>>>>>> scaler <<<<<<< HEAD ======= >>>>>>> scaler

Metric	lstm	naive	xgboost	cnn
diff_of_means	0.525	0.718	0.758	-2.057
ratio_of_sd	0.952	0.909	0.899	0.957
amplitude_ratio_of_means	0.772	0.641	0.609	0.748
maximum_error	0.198	0.174	0.152	0.145
ks_mean_on_coarse_res_with_extremes	0.142	0.245	0.324	0.194
qqplot_mae	0.057	0.115	0.106	0.105
acf_mae	0.078	0.118	0.130	0.092
0.500	0.718	0.758	-1.629
ratio_of_sd	0.962	0.909	0.899	0.978
amplitude_ratio_of_means	0.779	0.641	0.609	0.817
maximum_error	0.176	0.174	0.152	0.213
ks_mean_on_coarse_res_with_extremes	0.109	0.245	0.324	0.175
qqplot_mae	0.072	0.115	0.106	0.079
acf_mae	0.087	0.118	0.130	0.079
extremogram_mae	0.013	0.044	0.041	0.023
0.015

Time series of the first days

Plots

Important: Right now we are only estimating the upper tail extremogram. Currently we didn’t find a way to estimate the two tales at the same time. We are using quant = .97

<<<<<<< HEAD

=======

>>>>>>> scaler

cesm2.ssp585

<<<<<<< HEAD

<<<<<<< HEAD ======= >>>>>>> scaler <<<<<<< HEAD ======= >>>>>>> scaler

Metric	lstm	cnn	cnn	lstm	xgboost	naive
diff_of_means	0.972	-1.559	1.649	1.685
ratio_of_sd	0.952	0.960	0.891	0.897
amplitude_ratio_of_means	0.783	0.763	0.610	0.643
maximum_error	0.201	0.161	0.146	0.163
ks_mean_on_coarse_res_with_extremes	0.130	0.147	0.310	0.279
qqplot_mae	0.052	0.084	0.107	0.118
acf_mae	0.068	0.082	0.122	0.111
extremogram_mae	0.038	0.045	0.071	0.066
-0.885	1.373	1.649	1.685
ratio_of_sd	0.969	0.955	0.891	0.897
amplitude_ratio_of_means	0.829	0.786	0.610	0.643
maximum_error	0.218	0.173	0.146	0.163
ks_mean_on_coarse_res_with_extremes	0.151	0.126	0.310	0.279
qqplot_mae	0.062	0.068	0.107	0.118
acf_mae	0.068	0.075	0.122	0.111
extremogram_mae	0.037	0.041	0.071	0.066

Time series of the first days

Plots

Important: Right now we are only estimating the upper tail extremogram. Currently we didn’t find a way to estimate the two tales at the same time. We are using quant = .97

<<<<<<< HEAD

=======

>>>>>>> scaler

ec_earth3.ssp434

<<<<<<< HEAD

<<<<<<< HEAD ======= >>>>>>> scaler <<<<<<< HEAD ======= >>>>>>> scaler

Metric	lstm	cnn	xgboost	naive	xgboost	naive	lstm	cnn
diff_of_means	0.925	-1.538	-4.327	-4.509
ratio_of_sd	0.954	0.968	0.971	0.990
amplitude_ratio_of_means	0.766	0.743	0.638	0.665
maximum_error	0.195	0.137	0.180	0.135
ks_mean_on_coarse_res_with_extremes	0.257	0.273	0.405	0.388
qqplot_mae	0.101	0.080	0.173	0.178
acf_mae	0.085	0.098	0.132	0.128
extremogram_mae	0.094	0.098	0.135	0.131
-4.327	-4.509	-4.754	-6.878
ratio_of_sd	0.971	0.990	1.029	1.055
amplitude_ratio_of_means	0.638	0.665	0.804	0.848
maximum_error	0.180	0.135	0.201	0.248
ks_mean_on_coarse_res_with_extremes	0.405	0.388	0.274	0.263
qqplot_mae	0.173	0.178	0.188	0.272
acf_mae	0.132	0.128	0.094	0.095
extremogram_mae	0.135	0.131	0.093	0.093

Time series of the first days

Plots

Important: Right now we are only estimating the upper tail extremogram. Currently we didn’t find a way to estimate the two tales at the same time. We are using quant = .97

<<<<<<< HEAD

=======

>>>>>>> scaler

mri_esm2_0.ssp245

<<<<<<< HEAD

<<<<<<< HEAD ======= >>>>>>> scaler <<<<<<< HEAD ======= >>>>>>> scaler

Metric	lstm	cnn	xgboost	naive	cnn	xgboost	naive	lstm
diff_of_means	0.808	-1.506	36.022	36.479
ratio_of_sd	0.959	0.970	0.546	0.610
amplitude_ratio_of_means	0.764	0.750	0.368	0.522
maximum_error	0.202	0.163	0.269	0.239
ks_mean_on_coarse_res_with_extremes	0.194	0.224	0.373	0.233
qqplot_mae	0.065	0.066	1.425	1.442
acf_mae	0.083	0.097	0.128	0.084
extremogram_mae	0.030	0.036	0.045	0.026
32.753	36.022	36.479	38.667
ratio_of_sd	0.641	0.546	0.610	0.671
amplitude_ratio_of_means	0.645	0.368	0.522	0.607
maximum_error	0.166	0.269	0.239	0.196
ks_mean_on_coarse_res_with_extremes	0.062	0.373	0.233	0.040
qqplot_mae	1.295	1.425	1.442	1.529
acf_mae	0.031	0.128	0.084	0.049
extremogram_mae	0.027	0.045	0.026	0.009

Time series of the first days

Plots

Important: Right now we are only estimating the upper tail extremogram. Currently we didn’t find a way to estimate the two tales at the same time. We are using quant = .97

<<<<<<< HEAD

=======

>>>>>>> scaler

mri_esm2_0.ssp370

<<<<<<< HEAD

<<<<<<< HEAD ======= >>>>>>> scaler <<<<<<< HEAD ======= >>>>>>> scaler

Metric	lstm	cnn	xgboost	naive	cnn	xgboost	naive	lstm
diff_of_means	0.909	-1.383	36.673	37.285
ratio_of_sd	0.949	0.960	0.529	0.599
amplitude_ratio_of_means	0.769	0.756	0.361	0.526
maximum_error	0.183	0.156	0.273	0.213
ks_mean_on_coarse_res_with_extremes	0.144	0.183	0.323	0.204
qqplot_mae	0.057	0.073	1.451	1.474
acf_mae	0.065	0.079	0.115	0.067
extremogram_mae	0.011	0.017	0.032	0.019
33.020	36.673	37.285	39.359
ratio_of_sd	0.631	0.529	0.599	0.658
amplitude_ratio_of_means	0.647	0.361	0.526	0.604
maximum_error	0.177	0.273	0.213	0.225
ks_mean_on_coarse_res_with_extremes	0.043	0.323	0.204	0.102
qqplot_mae	1.305	1.451	1.474	1.556
acf_mae	0.045	0.115	0.067	0.040
extremogram_mae	0.030	0.032	0.019	0.028

Time series of the first days

Plots

Important: Right now we are only estimating the upper tail extremogram. Currently we didn’t find a way to estimate the two tales at the same time. We are using quant = .97

<<<<<<< HEAD

=======

>>>>>>> scaler

mri_esm2_0.ssp434

<<<<<<< HEAD

<<<<<<< HEAD ======= >>>>>>> scaler <<<<<<< HEAD ======= >>>>>>> scaler

Metric	lstm	cnn	xgboost	naive	cnn	xgboost	naive	lstm
diff_of_means	0.433	-1.964	35.896	36.372
ratio_of_sd	0.948	0.954	0.543	0.607
amplitude_ratio_of_means	0.767	0.748	0.371	0.525
maximum_error	0.205	0.151	0.278	0.227
ks_mean_on_coarse_res_with_extremes	0.170	0.189	0.357	0.221
qqplot_mae	0.061	0.090	1.420	1.438
acf_mae	0.078	0.094	0.125	0.084
extremogram_mae	0.014	0.014	0.027	0.027
32.468	35.896	36.372	38.496
ratio_of_sd	0.646	0.543	0.607	0.674
amplitude_ratio_of_means	0.657	0.371	0.525	0.624
maximum_error	0.161	0.278	0.227	0.217
ks_mean_on_coarse_res_with_extremes	0.074	0.357	0.221	0.089
qqplot_mae	1.284	1.420	1.438	1.522
acf_mae	0.043	0.125	0.084	0.047
extremogram_mae	0.052	0.027	0.027	0.036

Time series of the first days

Plots

Important: Right now we are only estimating the upper tail extremogram. Currently we didn’t find a way to estimate the two tales at the same time. We are using quant = .97

<<<<<<< HEAD

=======

>>>>>>> scaler